GIBBSTHUR

GIBBSTHUR is a software tool that facilitates the analysis and breeding evaluation of ranking traits in equine populations. It is a task known for its complexity due to the discontinuous nature of competition results and the influence of competitors' performances. The tool adopts the Thurstonian model, which assumes an underlying Gaussian liability for a horse's performance, to address the challenges of evaluating ranking traits. By utilizing Monte Carlo Markov Chain (McMC) techniques and a Gibbs Sampler scheme, including a data-augmentation step for the liability associated with ranking traits, GIBBSTHUR provides a sophisticated approach to analyzing both ranking and other continuous or threshold traits.

The software is capable of estimating variance and covariance components, predicting breeding values, and determining the average performance of competitors in events. Demonstrated through a simple example, GIBBSTHUR has proven its efficacy in accurately recovering simulated variance and covariance components. It also shows a strong correlation between simulated and predicted breeding values and between estimates of event effects and competitors' average additive genetic effect, thereby confirming its utility in generating predictions critical for breeding decisions.